https://github.com/kpriyanshu256/humanactivityrecognition
https://github.com/kpriyanshu256/humanactivityrecognition
activity-recognition ml pca sklearn svm
Last synced: about 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/kpriyanshu256/humanactivityrecognition
- Owner: kpriyanshu256
- Created: 2018-05-14T14:02:37.000Z (about 8 years ago)
- Default Branch: master
- Last Pushed: 2018-05-14T14:44:48.000Z (about 8 years ago)
- Last Synced: 2025-03-22T01:26:49.943Z (about 1 year ago)
- Topics: activity-recognition, ml, pca, sklearn, svm
- Language: Python
- Size: 54.7 KB
- Stars: 2
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# HumanActivityRecognition
## Objective
To classify common human activities like walking,standing,laying on the basis of readings obtained from smartphone sensors
## Dataset
Source: UCI ML Repository
Human Activity Recognition Using Smartphones Data Set
https://archive.ics.uci.edu/ml/datasets/Human+Activity+Recognition+Using+Smartphones#
## Model
* Dataset has 561 attributes so Principal Component Analysis(PCA) is used to reduce the dimension.
* Best results are obtained by taking about 200 principal components.
* Linear SVM("one vs one") was used to classify the data
## About Repository
* ActivityRecognition.py --- script to pickle data
* ActivityRecognition2.py --- classification script
* HAR pca.png -- image showing 2 principal components of the data
* TDT.png -- shows a plot of training, development and testing accuracies over number of principal components
## Results
* Training accuracy ~ 99%
* Development or cross-validation accuracy ~ 98%
* Testing accuracy ~ 95-96%
* Most mis-classifications were obtained for standing and sitting classes as there is not quite of a difference between the 2 postures.
## Future Scope
* Neural networks can be tried for the dataset